Rookout, a provider of a software-as-a-service (SaaS) platform designed to simplify the debugging of applications, has extended the reach of its tools into on-premises IT environments.
Company CTO Liran Haimovitch said its new Data On-Prem capability extends the reach of a bytecode manipulation capability developed by Rookout to insert a snapshot capture of a specific line of code to an on-premises IT environment. This eliminates the need for DevOps teams to upload code and associated data to the Rookout platform manually.
The Data On-Prem capability is made available via a software development kit (SDK) provided by Rookout, he said.
Legacy approaches to debugging code are essentially broken because the processes employed to fix code are too cumbersome, said Haimovitch. Rookout is designed to collect full-stack data without requiring code to stop running or impact code execution. That approach allows developers to debug live code on the Rookout platform and then reinsert the code into their applications.
Haimovitch said Rookout is making a case for transforming how application debugging is addressed within a DevOps process. Rather than relying on tools that need to be tightly integrated and maintained within a continuous integration/continuous deployment (CI/CD) pipeline, it will be easier for developers to debug live code on the Rookout platform, said Haimovitch.
That approach will enable DevOps teams to go well beyond simple observability to embrace true understandability of how applications work down to individual lines of code, he added.
Debugging applications has always been problematic. In an ideal world, developers would have a much better handle on what bugs need to be fixed earlier rather than later. By making it easier to debug live code, the number of issues that developers could address before an application is deployed in a production environment should increase. It may never be possible to address every issue before an application is deployed; however, the number of critical issues theoretically should decline. Many code issues go unaddressed simply because the effort required to debug that code is too great. Many developers too often put off fixing those issues until the next update usually because they are rushing to meet a delivery deadline.
At the very least, developers should repeat the same mistakes less often as they work to debug live code.
Regardless of how the debugging issue is addressed, issues within code in the age of microservices are starting to have a bigger impact. An issue within a line of code in a single microservice can have a cascading impact on all other elements of an application that depend on that microservice.
Debugging application code is never going to be enjoyable. The challenge is finding a way to make the entire process less painful. There may come a day soon when machine and deep learning algorithms help automate much of the debugging process. In the meantime, DevOps teams might want to re-evaluate existing approaches to debugging code that clearly leave much to be desired.
